Last updated: November 15, 2026
Sift, Signifyd and Riskified are three of the most common fraud prevention platforms for online businesses, but they work in different ways. Signifyd and Riskified approve orders and guarantee them, so they pay you back if an approved order turns into a fraud chargeback. Sift gives your team machine learning scores and rules, and you make and own the decisions. The short answer: choose Signifyd or Riskified if you want fraud liability off your books, and choose Sift if you have a fraud team, see fraud beyond the checkout, or want full control.
| Sift | Signifyd | Riskified | |
|---|---|---|---|
| Model | Self-managed scores and rules | Guaranteed decisions | Guaranteed decisions |
| Who owns fraud losses | You | Signifyd, for approved orders | Riskified, for approved orders |
| Fraud coverage | Payments, account takeover, signups, content and promo abuse | Payments, account protection, returns and claims abuse | Payments, account protection, policy abuse |
| Disputes | Dispute management add-on | Chargeback recovery included for guaranteed orders | Dispute Resolve for non-fraud chargebacks |
| Typical pricing | Platform fee or per event | Share of approved order value | Share of approved order value |
| Best for | Marketplaces, apps and teams with analysts | Mid-market and enterprise retail | High-ticket ecommerce, travel, ticketing |
| Headquarters | San Francisco, CA | San Jose, CA | New York, NY |
Sift scores events across the customer journey, including signups, logins, payments, content posts and promotions, using machine learning trained on a large global network. Your team sets thresholds and writes rules, then sends risky events to review queues. Because Sift covers more than checkout, it is popular with marketplaces, food delivery, travel and fintech apps where fraud also shows up as fake accounts, account takeover and scam listings. The trade-off is that you keep the chargeback liability and need people to manage the system.
Signifyd reviews each ecommerce order and returns an approve or decline decision. If it approves an order that later becomes a fraud chargeback, Signifyd reimburses the merchant. It also offers account protection, abuse prevention for returns and item-not-received claims, and payment optimization tools. Signifyd suits retailers that want a predictable cost and do not want to staff a manual review team.
Riskified also guarantees approved orders, and it is often chosen by merchants where false declines are expensive, such as electronics, luxury fashion, travel and event tickets. Its Policy Protect product targets refund and returns abuse, and Dispute Resolve automates responses to non-fraud chargebacks. Riskified is listed on the NYSE.
You run a marketplace or app with fraud in many places, you have analysts who want to tune rules, or you would rather keep the margin a guarantee provider charges.
You are a mid-market or enterprise retailer that wants guaranteed protection with broad platform integrations and abuse coverage.
You sell high-ticket goods or travel, false declines are costing you sales, and policy abuse is a growing problem.
In every case, ask each vendor to run a retro analysis on your past orders. Approval rate differences of even a point or two can matter more than the fee. For more options, see our list of Signifyd alternatives and our ranking of ecommerce fraud prevention tools.
Neither is better for every merchant. Both guarantee approved orders. Riskified is often favored for high-ticket and travel merchants, Signifyd across mid-market retail. Compare them on your own order data.
No. Sift is a self-managed platform: it provides scores, rules and workflows, and the merchant owns the decisions and the fraud losses.
Both typically charge a percentage of approved order value, which varies by vertical, volume and risk. Sift is usually priced as a platform or usage fee. See our guide on fraud prevention software costs.
Sift is the control option, while Signifyd and Riskified are guarantee options. Pick the model first, then compare vendors inside it on approval rates, coverage and integration fit. For more on the guarantee model, read what a chargeback guarantee is and whether it is worth it.